I have to admit, there is something pretty satisfying about asking AI a question and getting an answer almost immediately.

Need ideas for a training activity? Done. Need help organizing information? Done. Need a starting point for something you have been staring at for twenty minutes? AI probably has a suggestion before you can finish your coffee.

And I am not going to pretend that isn’t helpful. It is.

But this week’s readings made me wonder about something. Just because AI helps us finish something faster, does that mean it helps us think better?

Or are we sometimes skipping the most important part?

The Part Between the Question and the Answer

In Let’s Not Make AI the “Easy Button,” Fajardo (2025) makes an interesting point about creativity. The finished product isn’t necessarily where all the value is. There is also value in figuring things out, making choices, getting stuck, and deciding what matters.

That caught my attention because I think it is easy to confuse getting something done with actually learning something.

As a student, I understand the temptation to take the easier route. Between school, work, and everything else, having a tool that can help you move forward is a good thing.

But I also know that some of the moments when I have learned the most were the moments when I had to sit with an idea a little longer.

Sometimes the struggle is part of the learning.

Rosenfeld and Fizel (2026) make a similar argument in their discussion of AI and design research. They suggest that AI doesn’t always have to help us move faster. Sometimes it can help us slow down, notice something we missed, or challenge an assumption.

I hadn’t really thought about using AI that way before.

What if, instead of asking AI to give me the answer, I asked it what I might be overlooking?

AI Can Give Me Ideas. I Still Have to Decide What Matters.

This reminded me of our Articles of Serendipity activity.

I used AI to explore different connections related to training new supervisors who might complete a course but still feel uncomfortable having difficult conversations with employees.

AI came up with several possibilities. Some were interesting, and some weren’t really where I wanted to go.

But one connection stood out to me because of my experience working in harm reduction: meeting people where they are.

That is something I already understood from my own work, but I hadn’t immediately connected it to supervisor training.

It made me think about how we sometimes expect people to leave a training feeling completely ready to handle a difficult situation. In reality, they might understand the information and still need support, practice, or a smaller first step.

AI helped me explore the connection, but it didn’t know why that idea mattered to me. My own experience is what made it meaningful.

I think that is an important distinction.

Fajardo’s Articles of Serendipity activity is about following curiosity rather than expecting AI to hand us a perfect final product (Stanford d.school, n.d.). Looking back, that is exactly what made the activity useful. The interesting part wasn’t everything AI generated. It was noticing the one idea that made me stop and think.

A Good-Looking Solution Isn’t Always a Good Solution

This also connects to Boller and Fletcher’s (2020) discussion of prototyping.

They emphasize testing ideas with actual learners before investing all our time in building a finished training solution.

That seems especially important now that AI can help us create training materials so quickly.

Imagine asking AI to build an outline for a supervisor training course. It might produce something organized, professional-looking, and full of activities.

But what if the supervisors don’t need another course?

What if they need someone to practice a difficult conversation with them? What if they need feedback? What if the real problem is that they are afraid of saying the wrong thing?

A polished training outline won’t answer those questions.

We still have to talk to people, test ideas, and be willing to change our minds.

The Chat About Chat activity offers another interesting example. It suggests using AI to explore different interpretations of a conversation, but it also reminds us that AI’s interpretation might be wrong (Stanford d.school, n.d.).

I like that idea because it puts the responsibility back on us. AI can suggest another perspective, but it cannot tell us with certainty what someone else is thinking or feeling.

And when we are designing learning experiences for real people, that difference matters.

Maybe the Better Question Is What Happens Next

I don’t think the answer is to stop using AI. I use it, and I can see how valuable it can be.

But I am starting to think I need to pay more attention to how I use it.

Am I asking it to finish my thinking for me?

Or am I using it to give myself something new to think about?

There is a difference between accepting the first answer and asking, “What else could be going on here?”

There is also a difference between generating a training solution and finding out whether that solution will actually help someone.

I think that is what I am taking away from these readings. AI can make the process easier, but easier shouldn’t automatically be the goal. Sometimes we need the questions, the uncertainty, and even the wrong ideas to get somewhere worthwhile.

Maybe the best thing AI can give us isn’t a faster answer. Maybe it’s a question we wouldn’t have thought to ask on our own.

And maybe what we do with that question is where the real learning begins.

References

Boller, S., & Fletcher, L. (2020). Design thinking for training and development: Creating learning journeys that get results. ATD Press.

Fajardo, G. (2025, June 12). Let’s not make AI the “easy button”: Creativity in the age of AI, part one. Stanford d.school.

Rosenfeld, H., & Fizel, J. (2026, August 13). A human-centered framework for using AI in design research. IDEO.

Stanford d.school. (n.d.). Articles of serendipity.

Stanford d.school. (n.d.). Chat about chat.